• biodiversity assessment;
  • bryophytes;
  • Fagus sylvatica;
  • forest;
  • indicator;
  • Picea abies;
  • vascular plants


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Literature Cited

In this study, we present an approach for the identification of indicators for biodiversity within forest ecosystems. We analyze the data of stands of pure Norway spruce (Picea abies (L.) Karst) and European beech (Fagus sylvatica L.), as well as mixed P. abies–F. sylvatica forests in the Solling mountains (NW Germany). The analysis is based on 683 vegetation samples in total. For different plant groups, that is, vascular plants, bryophytes, Red List species, we investigate species numbers as a parameter of biodiversity. Species numbers are differentiated into three classes to describe low to high diversity. Plots are separately examined for the three different forest types. In order to take the species–area relationship into account, we only use relevés with a plot size of 100 m2. Our approach focuses on the probability to be in a defined range of species numbers, that is, class, if a certain species occurs. For the purpose of facilitating the differentiation of the classes, we use the presence values of species in the classes for further characterization of indicators. Few indicators were found for the low ranges of species numbers. In addition, there were only a small number of species groups and stand types having indicators for all three classes. Various species have multiple indicator functions, e.g., with regard to the investigated species groups. The focus on a few of these multi-indicators allows a rapid assessment of forest biodiversity. The catalog of indicators resulting from the investigation helps to facilitate and accelerate biodiversity evaluations of forest stands, in particular with regard to nature conservation and the restoration of natural forests.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Literature Cited

The definition of biodiversity in article 2 of the Convention on Biological Diversity takes the different levels of organization and interactions of living organisms into account and thus shows that there is no single diversity. Up to now, different definitions have been developed in order to make the term more practicable for nature conservation and other purposes (Beierkuhnlein 1998; Kratochwil 1999). It turns out that biodiversity can never be entirely described, but some of its features can be measured (Mayer et al. 2002). Its complexity causes problems of measurement in practice (e.g., Levin 1997; Wiegleb 2003). Despite the existence of various indices, the most easily communicated approach is species count (e.g., Heywood et al. 1995).

Vascular plants are considered a key species group in forest ecosystems because of their contribution to primary production (Mitchell & Kirby 1989; Barthlott et al. 2000; Haeupler 2000). In addition, the knowledge about their taxonomy, ecology, and spatial distribution is relatively extensive, at least for Central Europe. Furthermore, there are studies that suggest a positive correlation between plant species number and variety in animal species composition (Andow 1991; Gaston 1992). Otherwise, various papers give examples, where stand types with species-poor vegetation are habitat for a diverse fauna (Kratochwil 1999; Pärt & Söderström 1999).

The identification of biodiversity indicators is an important basis for practical purposes like nature conservation, natural resources management, and ecosystem restoration. Additionally, indicator species have a long tradition for the assessment and monitoring of ecosystems (Noss 1999). Several requirements for indicators are identified (Cook 1976; Sheehan 1984; Munn 1988), but usually they are not all fulfilled by a single indicator. The demands applicable to our approach are a wide distribution of the indicator species and an easy and cost-effective detection.

There are various studies trying to predict plant species richness by or with the help of indicator species (e.g., Dufrêne & Legendre 1997; Dumortier et al. 2002). Many of the already developed methods for identification of indicators suffer from methodological problems (Clarke 1993; Belbin & McDonald 1993). For example, the use of relative abundances leads to distortion in classification procedures (Jackson 1997). Problems can especially be due to the use of data recorded by different authors and relatively rough information on abundances. Vane-Wright (1996) points out the risks of using rare species as indicators. For instance, it could be possible to choose unsteady species or individuals from sink populations. According to McGeoch and Chown (1998), an indicator would be most valuable if it would be determined independently of other species, would occur only in one of the site-related groups (that have been preliminarily defined in accord to the surveyed question), and would have a high frequency and abundance on those sites.

The present study analyzes plant species of the moss and herb layer in forests as possible indicators and measures the aspect of biodiversity by species number per plot. We also examine properties of the objects such as endangerment or affiliation to groups, such as herb or woody species or bryophytes. We hypothesize that (1) there are plant species among the total species pool, which predict defined ranges of species numbers, and (2) that they can be detected by our approach. The proposed approach is not supposed to substitute comprehensive surveys of vegetation. However, selected indicators allow a rough scan of forest stands with regard to the probability of finding defined ranges of species numbers. These scans can be carried out with a knowledge focused on the selected indicators and can precede extensive surveys accomplished by specialists in botany.

Additionally, biodiversity indicators are necessary for the evaluation of forests for practical purposes, such as nature conservation or forest restoration (e.g., conversion from anthropogenic toward natural forests). The possible reuse of data collected in the context of other ecological studies is not only another possibility to save time and other resources but could also be stimulation for scientists to work toward further standardization in data collection.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Literature Cited

Site Description

The study is based on vegetation data from forests of the Solling, a mountain range in Niedersachsen (NW Germany) that rises up to 528 m above sea level. The area is characterized by a relatively low mean annual temperature of approximately 7°C, low fluctuations between temperature extremes and a humid climate, with a mean annual precipitation around 1,000 mm (Deutscher Wetterdienst 1964). The Solling is a red sandstone plateau consistently covered by loess. Therefore, wide areas have acid silty loam cambisols (brown soil according to Ellenberg et al. 1986 and NMELF 1996).

The present-day vegetation cover is the consequence of long-lasting human impact. Until the ninth century, the Solling was largely covered by natural oligotrophic beech forests (Gerlach 1970). Since the Middle Ages, intensification of forest use (e.g., charcoal burning, forest grazing, and forest glass works) has led to the degradation of forests and forest sites. Consequently, Norway spruce (Picea abies (L.) Karst), which did not occur naturally in the Solling, was introduced in the early 1700s for the reforestation of the degraded forest sites. At present, the forests of the Solling are still dominated by spruce plantations. Additionally, beech, mixed beech–spruce, and oak forests exist (Ellenberg et al. 1986; Zerbe 1992; NMELF 1996).


The analyzed database contains data from two studies on forest ecosystems in the Solling, including in total 683 vegetation samples from Gerlach (1970) and Weckesser (2003). The vegetation samples with a plot size of 100 m2 provide information on frequency and abundance of plant species in canopy and understory vegetation. Three stand types are examined, that is, spruce (P. abies (L.) Karst), beech (Fagus sylvatica L.), and mixed forests with beech and spruce. Stands are defined as mixed forests if both tree species occur in the canopy of the forest. In the pure stands, only beech or only spruce makes up the canopy. Abundance data are given according to the scale of Braun-Blanquet (1964) or in steps of 1 to 5%. The different species synonyms are unified according to Wisskirchen and Haeupler (1998) for vascular plants and Frahm and Frey (2004) for bryophytes. Common names are derived from the plants database of the USDA/NRCS (2004) or, if data on species are lacking, from BSBI (2004).

Species number is used as a feature of biodiversity and analyzed with regard to different groups of plants (1Fig. 1). Below, we refer to the species numbers of the plant groups that can be predicted by an indicator also as “indicator functions.” The group of woody plants includes phanerophytes (trees), nanophanerophytes (shrubs), and biennial nanophanerophytes (Ellenberg et al. 1992), e.g., Shrubby blackberry (Rubus fruticosus agg.), in the herb, shrub, and tree layer. Grass-like species include the families Poaceae, Cyperaceae, and Juncaceae. Information on endangered species was taken from the regional (federal states of Niedersachsen and Hessen) and national Red Lists compiled by Ludwig and Schnittler (1996) as well as the regional Red List of the federal state of Nordrhein-Westfalen (LÖBF 1999). Species assigned to categories 1 (critically endangered) through 3 (vulnerable), as well as species that belong to category V (decreasing frequency), add up to the number of endangered species.


Figure 1. Plant groups, for which biodiversity indicators were examined. Relationships between the species groups (e.g., overlapping) are displayed qualitatively, not quantitatively. Our data have no intersection between woody and endangered species.

Download figure to PowerPoint

We check the data for stand-type characteristic differences in ranges of species numbers as well as for the influence of the canopy dominants, beech and spruce. The test for significance that is used is a median test (Woolson 1987).

Definition of Species Number Classes

Our approach uses probabilities of predefined ranges of species numbers to occur, if certain species of higher plants and bryophytes are found. For every analyzed species group we predefine three classes for the ranges of species numbers. The class 1 includes the species-poor stands, class 2 the stands with intermediate species numbers, and class 3 comprises the species-rich stands (1Table 1). We strive for equal sample sizes in all three classes.

Table 1.  Species number of plant groups in stand types and classes.
Moss and herb layer0–56–89–171–1213–1920–407–1718–2728–47
Herb layer0–34–56–140–78–1314–313–1011–1819–36
Herb species012–60–23–56–171–45–1011–23
Grass species01–23–80–12–34–110–23–45–10
Woody species123–51–234–71–23–45–8
Endangered species012–4

Steps for Selecting Indicators

There are three steps for the selection of indicator species from the moss, herb, and shrub layer (2Fig. 2). The first step and crucial criterion is that species xi should be an indicator if probability p(cj|xi) for a class cj is higher than for the other two classes. The probability to be in class cj if we have observed species xi is p(cj|xi), which means the probability to be in a site with a species number within the range defined for cj. For reasons of practicability, we demand of an indicator for class j that p(cj|xi) ≥ 0.6, that is, xi may be an indicator for class cj if at least 60% of the plots where we find xi have a species number within the defined range for this class. The aspect of practicability is also motivation for the second selection step, which requires that the highest probability p(xi|cj) to find species xi given that we are sampling in class cj has to be in the same class as the maximum p(ck|xi), with j=k. The probability p(xi|cj) corresponds to the degree of species presence. If the maxima of the probabilities p(ck|xi) and p(xi|cj) are not in the same classes with kj, the species xi is rejected as an indicator. In this way, we avoid using a species as an indicator that is more frequent in another class of species numbers than in the class it is an indicator for. The described case with kj can only happen if the numbers of plots in the classes are not equal. The third selection step filters the species regarding frequencies of at least 5% among all plots.


Figure 2. Methodological steps for the identification of biodiversity indicators.

Download figure to PowerPoint

Additionally, differences in percent cover of indicators between the classes are tested. A higher abundance of an identified indicator species in the indicated class supports the practicability of the surrogate.

To summarize the approach, we can say that the decisive property for an indicator is the ability to predict a defined range of species numbers cj with a probability p(cj|xi) greater than p(ck|xi), kj. The other selection steps facilitate the use of the indicator, avoid the choice of rare species, and are therefore necessary for the practicability of the indicator species.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Literature Cited

Dependence on Canopy Dominants

Of all species recorded in the moss, herb, and shrub layer of the forest stands, 28 belong to woody, 33 to grass, and 54 to herb species, as well as 45 to bryophytes. Among the woody species, 12 occur, additionally to the herb layer, in the shrub layer. Interrupted clubmoss (Lycopodium annotinum), Bilberry (Vaccinium myrtillus), and Heather (Calluna vulgaris) are not assigned to the categories woody, grass, and herb species. Thirteen of the 163 species recorded are considered to be endangered. For species taken into consideration, occurrences in herb and shrub layer are treated like separate species (2Table 2).

Table 2.  Class prediction by indicators in examined plant groups and stand types.
  1. p100=p(cj I xi) × 100; DP, degree of presence [in %]; S, shrub layer; HL, herb layer; H, herb species; G, grass species; B, moss layer/bryophytes; W, woody species; E, endangered species.

Moss and herb layer
 Acer pseudoplatanusHLW 36310
 Agrostis capillarisHLG 3904536661
 Athyrium filix-feminaHLH 3615137769
 Atrichum undulatumBB 2601638557
 Brachythecium velutinumBB 37741
 Calamagrostis epigejosHLG 3682239039
 Campylopus flexuosusBBE26231
 Cardamine flexuosaHLH 373203613736349
 Carex ovalisHLG 370183711937124
 Carex pallescensHLG 3781836714
 Carex remotaHLG 36817
 Deschampsia cespitosaHLG 310020
 Dicranum scopariumBB 26016
 Digitalis purpureaHLH 3781736059
 Epilobium angustifoliumHLH 36754
 Epilobium ciliatumHLH 38814
 Epilobium montanumHLH 3771838463
 Eurhynchium praelongumBB 36355
 Fagus sylvaticaHLW 36563
 Fagus sylvaticaSW 38033
 Festuca giganteaHLG 39224
 Galeopsis bifida et tetrahitHLH 3641436843
 Galium aparineHLH 310049
 Galium saxatileHLH 36372
 Gymnocarpium dryopterisHLH 3681737316
 Holcus lanatusHLG 310016
 Holcus mollisHLG 3862938135
 Impatiens noli-tangereHLH 39222
 Impatiens parvifloraHLH 38476
 Isopterygium elegansBB 39435
 Juncus effususHLG 376483674436753
 Moehringia trinerviaHLH 38135
 Mycelis muralisHLH 3711337680
 Oreopteris limbospermaHLH 310018
 Picea abiesHLW 36338
 Plagiothecium undulatumBBE37120
 Poa trivialisHLG 3871639224
 Pohlia nutansBB 37814
 Quercus petraea et roburHLW 36012
 Ranunculus repensHLH 39222
 Rubus fruticosus agg.HLW 3762937567
 Salix capreaHLW 38016
 Senecio ovatusHLH 38727
 Sharpiella seligeriBB 36178
 Sorbus aucupariaSW 37922
 Stellaria alsineHLH 3861437633
 Stellaria mediaHLH 37721
 Taraxacum spp.HLH 36720
 Trientalis europaeaHLH 37413
 Urtica dioicaHLH 388183613336659
 Vaccinium myrtillusHL 36054
 Veronica officinalisHLH 39057
Herb layer
 Acer pseudoplatanusHLW 36310
 Agrostis capillarisHLG 31003837068
 Athyrium filix-feminaHLH 3100173645638071
 Atrichum undulatumBB 38255
 Brachythecium velutinumBB 37339
 Calamagrostis epigejosHLG 3802839039
 Cardamine flexuosaHLH 3911936349
 Carex canescensHLGE26712
 Carex ovalisHLG 390173762237124
 Carex pallescensHLG 37818
 Carex piluliferaHLG 36366
 Carex remotaHLG 3691736517
 Deschampsia cespitosaHLG 310020
 Deschampsia flexuosaHLG 36738
 Dicranum polysetumBBE16717
 Digitalis purpureaHLH 3851936059
 Epilobium angustifoliumHLH 36353
 Epilobium ciliatumHLH 38814
 Epilobium montanumHLH 3701838665
 Eurhynchium praelongumBB 36053
 Fagus sylvaticaHLW 36765
 Fagus sylvaticaSW 3644037531
 Festuca giganteaHLG 3621139224
 Galeopsis bifida et tetrahitHLH 3681636843
 Galium aparineHLH 310049
 Gymnocarpium dryopterisHLH 3711837316
 Holcus lanatusHLG 38814
 Holcus mollisHLG 3863138135
 Impatiens noli-tangereHLH 39222
 Impatiens parvifloraHLH 3731538476
 Isopterygium elegansBB 39435
 Juncus effususHLG 388423775336753
 Moehringia trinerviaHLH 38637
 Mycelis muralisHLH 3711437882
 Oreopteris limbospermaHLH 310018
 Picea abiesHLW 37936
 Picea abiesSW 26312
 Pleurozium schreberiBB 17015
 Poa trivialisHLG 3831638522
 Pohlia nutansBB 37814
 Quercus petraea et roburHLW 36012
 Ranunculus repensHLH 39222
 Rubus fruticosus agg.HLW 3712937365
 Rubus idaeusHLW 37466
 Salix capreaHLW 37014
 Senecio ovatusHLH 38727
 Sharpiella seligeriBB 36178
 Sorbus aucupariaSW 37120
 Stellaria alsineHLH 3811437633
 Stellaria mediaHLH 3712136253
 Taraxacum spp.HLH 36720
 Tetraphis pellucidaBB 16148
 Trientalis europaeaHLH 37414
 Urtica dioicaHLH 3881336659
 Veronica officinalisHLH 39459
Herb species
 Acer pseudoplatanusHLW 37513
 Agrostis capillarisHLG 36023
 Athyrium filix-feminaHLH 3891537772
 Atrichum undulatumBB 37955
 Brachythecium velutinumBB 36938
 Calamagrostis epigejosHLG 37132
 Campylopus flexuosusBBE16935
 Cardamine flexuosaHLH 3100213724637157
 Carex ovalisHLG 37126
 Deschampsia cespitosaHLG 37015
 Digitalis purpureaHLH 3851936364
 Dryopteris carthusianaHLH 38440
 Dryopteris dilatataHLH 37142
 Epilobium angustifoliumHLH 36353
 Epilobium ciliatumHLH 38815
 Epilobium montanumHLH 3902338970
 Fagus sylvaticaSW 3613836528
 Festuca giganteaHLG 39226
 Galeopsis bifida et tetrahitHLH 3751836845
 Galium aparineHLH 310051
 Gymnocarpium dryopterisHLH 3741937317
 Holcus lanatusHLG 36311
 Holcus mollisHLG 36730
 Impatiens noli-tangereHLH 310026
 Impatiens parvifloraHLH 3911938681
 Isopterygium elegansBB 37830
 Juncus effususHLG 36251
 Luzula pilosaHLG 26212
 Moehringia trinerviaHLH 38136
 Mycelis muralisHLH 3831737885
 Oreopteris limbospermaHLH 38917
 Oxalis acetosellaHLH 36275
 Pleurozium schreberiBB 16012
 Poa trivialisHLG 3741436919
 Ranunculus repensHLH 39223
 Rubus fruticosus agg.HLW 36728
 Rubus idaeusHLW 36860
 Senecio ovatusHLH 39330
 Stellaria alsineHLH 3901637634
 Stellaria mediaHLH 3862536760
 Taraxacum spp.HLH 36019
 Tetraphis pellucidaBB 16649
 Trientalis europaeaHLH 37815
 Urtica dioicaHLH 388133703936864
 Veronica officinalisHLH 39462
Grass species
 Acer pseudoplatanusHLW 26310
 Agrostis capillarisHLG 3100423846536493
 Athyrium filix-feminaHLH 378153704937557
 Atrichum undulatumBB 38548
 Brachythecium velutinumBB 38136
 Calamagrostis epigejosHLG 3902539534
 Cardamine flexuosaHLH 37317
 Carex ovalisHLG 390193942138826
 Carex pallescensHLG 37815310018
 Carex piluliferaHLG 37183
 Carex remotaHLG 3691938117
 Deschampsia cespitosaHLG 310017
 Deschampsia flexuosaHLG 37346
 Digitalis purpureaHLH 3781436352
 Epilobium angustifoliumHLH 36544
 Epilobium ciliatumHLH 3639
 Epilobium montanumHLH 3671337347
 Fagus sylvaticaHLW 37360
 Fagus sylvaticaSW 39031
 Festuca giganteaHLG 3761138519
 Galeopsis bifida et tetrahitHLH 36813
 Galium aparineHLH 39640
 Galium saxatileHLH 36563
 Gymnocarpium dryopterisHLH 3741536412
 Holcus lanatusHLG 310014
 Holcus mollisHLG 39828310036
 Impatiens noli-tangereHLH 38317
 Impatiens parvifloraHLH 37053
 Isopterygium elegansBB 39429
 Juncus effususHLG 396503894938557
 Lepidozia reptansBB 26221
 Luzula pilosaHLG 38515
 Moehringia trinerviaHLH 38631
 Molinia caeruleaHLG 37812
 Mycelis muralisHLH 36557
 Oreopteris limbospermaHLH 310016
 Oxalis acetosellaHLH 26054
 Picea abiesHLW 36331
 Plagiothecium undulatumBBE36014
 Poa trivialisHLG 39615310022
 Pohlia nutansBB 38914
 Ranunculus repensHLH 37516
 Rubus fruticosus agg.HLW 3672237053
 Rumex acetosellaHLH 36916
 Salix capreaHLW 39016
 Senecio ovatusHLH 37319
 Sorbus aucupariaSW 38621
 Stellaria alsineHLH 3811136724
 Stellaria mediaHLH 36615
 Taraxacum spp.HLH 36717
 Tetraphis pellucidaBB 27414
 Trientalis europaeaHLH 37411
 Urtica dioicaHLH 38815
 Vaccinium myrtillusHL 36146
 Veronica officinalisHLH 39450
 Acer pseudoplatanusHLW 37512
 Atrichum undulatumBB 3933037648
 Brachythecium rutabulumBB 36140
 Brachythecium velutinumBB 37738
 Calamagrostis epigejosHLG 37129
 Calypogeia muellerianaBB 36727
 Carex canescensHLGE17815
 Carex pallescensHLG 38917
 Carex remotaHLG 16213
 Deschampsia cespitosaHLG 37013
 Dicranum scopariumBB 37323
 Epilobium montanumHLH 36244
 Eurhynchium praelongumBB 37058
 Fagus sylvaticaSW 37027
 Festuca giganteaHLG 36917
 Galium aparineHLH 38338
 Gymnocarpium dryopterisHLH 36413
 Holcus lanatusHLG 36310
 Hypnum cupressiformeBB 38664
 Impatiens noli-tangereHLH 38319
 Impatiens parvifloraHLH 1731336152
 Isopterygium elegansBB 3852337827
 Lepidozia reptansBB 36232
 Moehringia trinerviaHLH 36727
 Molinia caeruleaHLG 16713
 Oreopteris limbospermaHLH 38915
 Plagiothecium undulatumBBE37429
 Poa trivialisHLG 36917
 Pohlia nutansBB 37813
 Ranunculus repensHLH 38319
 Sharpiella seligeriBB 36679
 Veronica officinalisHLH 36137
Woody species
 Acer pseudoplatanusHLW 2638
 Athyrium filix-feminaHLH 36044
 Atrichum undulatumBB 17315
 Calamagrostis epigejosHLG 3712037150
 Campylopus flexuosusBBE16933
 Carex canescensHLGE27811
 Carex ovalisHLG 3602636515
 Carex pallescensHLG 36713
 Deschampsia cespitosaHLG 36020
 Dicranum polysetumBBE17516
 Epilobium angustifoliumHLH 36848
 Epilobium montanumHLH 36013
 Eurhynchium praelongumBB 36125
 Fagus sylvaticaSW 3674638557
 Galeopsis bifida et tetrahitHLH 37915
 Gymnocarpium dryopterisHLH 26411
 Holcus lanatusHLG 310027
 Impatiens noli-tangereHLH 36727
 Impatiens parvifloraHLH 26413
 Juncus effususHLG 36135
 Luzula pilosaHLG 26010
 Molinia caeruleaHLG 26710
 Mycelis muralisHLH 36711
 Oreopteris limbospermaHLH 37823
 Pleurozium schreberiBB 17013
 Poa trivialisHLG 3741237733
 Quercus petraea et roburHLW 36020
 Rubus fruticosus agg.HLW 3943236697
 Rubus idaeusHLW 36991
 Rumex acetosellaHLH 26213
 Salix capreaHLW 36020
 Sorbus aucupariaSW 310047
 Stellaria alsineHLH 37110
 Stellaria mediaHLH 36015
 Tetraphis pellucidaBB 16847
 Urtica dioicaHLH 36128
Endangered species
 Atrichum undulatumBB16743
 Brachythecium velutinumBB16533
 Calamagrostis epigejosHLG17129
 Campylopus flexuosusBB38160
 Carex canescensHLG37820
 Deschampsia cespitosaHLG17014
 Dicranum polysetumBB36723
 Fagus sylvaticaSW18031
 Galium aparineHLH16329
 Holcus lanatusHLG18814
 Holcus mollisHLG17129
 Impatiens noli-tangereHLH18320
 Isopterygium elegansBB18931
 Luzula pilosaHLG27011
 Moehringia trinerviaHLH17129
 Oreopteris limbospermaHLH17814
 Poa trivialisHLG17720
 Pohlia nutansBB17814
 Sorbus aucupariaSW17120

Maximal species numbers range from 4 species for endangered species in spruce stands to 47 species in the moss and herb layer of spruce stands (Table 1). In general, the spruce forests reach the highest species numbers among the stand types. The analysis for endangered species is restricted to spruce stands due to the low species numbers in the beech and beech–spruce forests. Beech stands have the lowest ranges of species numbers for all examined species groups.

For woody species, we are not able to detect significant differences between the species numbers in mixed and spruce stands (median test, p > 0.05). The other species groups show significant differences between the medians of species numbers in all examined stand types. To check the general influence of beech and spruce in the canopy on species numbers, we use the chosen dataset and compare the species numbers of presence–absence groups for both tree species. Differences between species numbers with regard to all the seven analyzed groups are significant with higher medians generally in stands without beech or with spruce in the canopy (median test, p < 0.05).

Biodiversity Indicators

Based on our selection steps, we found 70 indicator species, meeting all three selection criteria, within the analyzed species groups (Table 2). Indicators are similarly distributed among the classes within the different species groups, e.g., most of the indicators predict class 3. With 45 indicators, spruce plots have the largest number of indicator species among the stand types for a certain species group.

The identified indicator species show different distributions along the gradient of species numbers, which is shown in 3Figure 3, taking Silvery sedge (Carex canescens), Stickywilly (Galium aparine), and the bryophyte Tetraphis pellucida as examples. Averaged for all three stand types, the selection steps for the identification of indicator species show similar consequences for the species groups with regard to the percentage of reduction of the precedent species pool (4Fig. 4).


Figure 3. Distribution of species numbers of the herb layer for spruce stands where three selected indicators occur.

Download figure to PowerPoint


Figure 4. Reduction of possible indicators by sequenced selection steps (Fig. 2) dependent on examined groups. Presented values are means of the stand types studied. Endangered species are not considered because they are only examined for spruce stands.

Download figure to PowerPoint

Concerning the affiliation to the examined groups, the largest fraction is derived from the herb layer (in total 51 indicators) with 16 grass, 27 herb, and 10 woody species. Sixteen indicators are bryophytes, and four indicators are considered endangered. Spruce forests have the highest numbers of indicators, followed by beech–spruce and beech stands. An exception is observed with regard to the bryophyte species, where beech–spruce stands have a lower number of indicators than pure beech stands (5Fig. 5). Most of the plant species limited in their function as indicator within one stand type are confined to the spruce plots (26). Indicators that are unique to beech (9) and mixed stands (7) are less frequent. A large number of indicators belong to the same group, which is accounted for by the species number calculation (i.e., intraindicators). This is especially true for indicators of class 3.


Figure 5. Number of indicators depending on stand type and examined group. MH, moss and herb layer; HL, herb layer; H, herb species; G, grass species; B, bryophytes; WS, woody species; ES, endangered species.

Download figure to PowerPoint

There are only four of the examined species groups within the stand types having indicators for all three species number classes: woody species in spruce and beech stands, herb layer in spruce stands, and endangered species in spruce stands. In particular, indicators for class 1 are often lacking. Indicators for a small number of woody species include Atrichum undulatum (p(c1|x) = 0.73, p(x|c1) = 0.15, p(x|c2) = 0.07, p(x|c3) = 0, frequency p(x) = 0.1) in beech forests and the mosses Campylopus flexuosus and Dicranum polysetum in spruce forests. A small number of endangered species in spruce stands is indicated by, inter alia, G. aparine and Western touch-me-not (Impatiens noli-tangere). Remote sedge (Carex remota) and Small balsam (I. parviflora) indicate bryophyte-poor beech stands.

Species-rich stands are best represented by indicators. Every examined group has indicators for class 3 in contrast to class 2. One indicator species fulfills up to six indicator functions for spruce stands (indicator function for endangered species excluded), in beech stands up to four, and in mixed forests up to five functions. The largest fraction of indicator species as listed in Table 2 has four indicator functions for spruce and beech–spruce stands and two to three indicator functions for beech forests. For example, in spruce stands, Common velvetgrass (Holcus lanatus) is an indicator for large species numbers among woody species, grasses, herbs, bryophytes, as well as plants in the herb layer and the moss and herb layer. With regard to the prediction certainty p(cj|xi), G. aparine, European mountain ash (Sorbus aucuparia, shrub layer), Common gypsyweed (Veronica officinalis), Creeping velvetgrass (H. mollis), Giant fescue (Festuca gigantea), and Rough bluegrass (Poa trivialis) are among the strongest indicators for all stand types, classes, and groups. If those species occur, the probability is on average greater than 0.8 to be in a defined class. Most indicator species show no significant difference in percent cover between the class for which they are an indicator and the other two classes. In contrast, Rubus fruticosus agg. as an indicator for a high species number of woody species in spruce stands has a higher abundance in class 3 (0.2%) than in the other two classes (0.01%), besides p(c3|x) = 0.66 and p(x|c3) = 0.97.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Literature Cited

Methodological Aspects

There are several advantages of our approach in comparison to the existing methods applied and reviewed, respectively, by Hill (1979), Dufrêne and Legendre (1997), and Dumortier et al. (2002). First, the method is not bound to certain site characterization methods and is not limited to ordination methods like TWINSPAN (Hill 1979). Ordination methods suffer from the restriction to a limited percentage of the variability that can be described and therefore make TWINSPAN most suitable for data with a strong and unidimensional underlying gradient (Belbin & McDonald 1993). This strong gradient cannot necessarily be expected for species occurrences and species numbers due to the fact that species numbers correspond to various abiotic and biotic influences. However, we use samples of relatively homogeneous stand types.

Second, we use presence–absence data for the indicator search, which minimizes distortion by different authors or sampling methods (Dufrêne & Legendre 1997). In Germany, percent cover is usually estimated according to the phytosociological method of Braun-Blanquet (1964) or its modifications (e.g., Bemmerlein-Lux et al. 1994). If data from different authors are analyzed, this can lead to distortion, assuming that the surveyors do not make exactly the same cover percentage estimation in the field. By considering only information on presence or absence of a species, we avoid this error and enable the use of already existing data from various studies and authors, respectively, resulting in cost-effective pilot surveys, especially in the well-examined areas of central Europe.

Our approach is not restricted to the plant groups examined here and can be extended to other sets (e.g., indigenous species or exotic species) or adjusted to other groups of plant taxa. By using other ranges of species richness, also the scope of the indicator can be further specified.

Dependence of Species Number on Plot Size and Stand Type

The dependence of species richness on plot size leads to the necessity to specify the spatial scale, for which results of our approach are valid. We assume that indicators found within the present study are valid for managed forests on acidic sites in central European low mountain ranges. This assumption is in accordance with Dufrêne and Legendre (1997) who state that the derived indicators can only be used in habitats that are similar to those used to find the indicator species. In addition, they should be applicable to plots with a size of about 100 m2.

As outlined above, stands have to be assessed depending on the stand type. It is well revealed that species numbers increase in order from beech to mixed to spruce forests in regard to different plant groups, e.g., vascular plants and bryophytes (Ellenberg et al. 1986; Zerbe 1992; Weckesser 2003). The increasing species numbers from near-natural forests to anthropogenic forest stands (Mooney et al. 1995 and Dierßen & Kiehl 2000) correspond well to the intermediate disturbance hypothesis (Grime 1973; Connell 1978; Hobbs & Huenneke 1992), whereas the observed ranking from spruce plantations to beech stands also depends on stand conditions such as the site factor light (cf Passarge 1968).

Whereas the number of recorded endangered species in beech and mixed beech–spruce forests is negligible, these species are relatively common in spruce forests. Higher numbers of endangered species, often observed in coniferous plantations (Trautmann 1976; Philippi et al. 1993; Zerbe 1999a), are the consequence of stand conditions (e.g., relatively low canopy cover and humus accumulation in old-growth stands) and/or former land use of the forest site such as pasture (Zerbe 1999b). In our study, mostly bryophytes (26) contribute to the total number of endangered species (38).

Distribution of Indicators among Stand Types, Plant Groups, and Classes

With regard to the distribution of indicators among the examined three classes, the under-representation of class 2 and, in particular, class 1 is conspicuous. There are species with p(cj|xi) ≥ 0.6 for class 1, but they are not among the detected indicators due to their low frequency. Other species occurring in the low and intermediate range of examined species numbers do not meet the selection requirements. Similar to the species numbers among the groups, which are highest for spruce stands and lowest for beech forests, the number of indicators is also maximal for spruce forests. An exception from this general tendency can be observed for bryophyte richness in beech–spruce forests, which is indicated by a smaller number of species than bryophytes in beech forests. However, the total number of bryophyte indicators is low due to the fact that the occurrence of bryophytes in beech forests is inhibited by leaf litter (Longton & Greene 1979; Økland 1995) and often restricted to sites mostly free of litter, e.g., trunk feet and tree stumps (Ellenberg 1996). Regarding bryophytes, mixed stands have lower species numbers in comparison to spruce stands but larger species numbers than beech forests. The restriction to special sites does not have the same effect as in beech forests due to the presence of spruce and the conditions favorable for bryophytes associated with spruce (Zerbe 1992). The shady conditions within beech forests are lowering the total species numbers in a plot. In summary, we have a larger range of suitable sites for bryophytes in mixed stands but a smaller number of possible indicators.

Many of the indicators detected by our approach indicate biodiversity of the plant group that they are part of (intraindicators, e.g., Holcus lanatus as an indicator for grass species richness). Various reasons are considered responsible for this. The number of species belonging to the group that is used for calculating species number classes in relation to the total species pool is very important for this observation. Taking the group of species in the moss and herb layer as an example, it is obvious that most of the indicator species have to be part of this group due to scarce alternatives. In this case, other indicators could only be derived from the shrub layer. The larger percentage of intraindicators in the species-rich class 3 can be explained by another reason. With increasing species number, the probability of finding a certain species from this group increases as well. Following this assumption, the frequency of that species is usually highest in class 3, which can lead to a detection of that species as an indicator following our approach. In addition, stands rich in species of one group are not necessarily species rich for another group, which again limits the pool of potential indicators. For example, the species number of bryophytes is not well represented by diversity hot spots of vascular plants (Dirkse & Martakis 1998).

In particular, classes with a large number of indicators contain several subgroups, which are similar with regard to their ecological range. This allows to summarize more specific ecological groups (e.g., species with a high light demand) than the whole indicator pool for a given species richness. Pooling indicator species does not increase the probability p(cj|xi), which can be demonstrated for Broad-leaved willowherb (Epilobium montanum), Impatiens parviflora, and Wall-lettuce (Mycelis muralis). The prediction probability p(c3|xi) ranges for these species between 0.76 and 0.84, whereas the pooled is only 0.71.

Our approach avoids rare species (Vane-Wright 1996) in order to facilitate the detection of indicator species in forest stands and therefore diminishes the need for summarization of indicator groups according to their ecological properties. But it would also be possible to use rare species as additional indicators in order to support the principal indicator species and increase p(cj|xi).

With regard to the robustness of biodiversity indicators, we assume that a “good” indicator would not only be more frequent in the indicated class following detection step 1, p(cj|xi) ≥ 0.6, but would also have a higher percent cover. The few species, which have a higher abundance in the indicated class than in the other two classes, usually show only small differences. For example, we have an abundance of 0.2% for the indicated class and an abundance of 0 to 0.01% for the other two classes. Consequently, the analyses of our data seem to be a hint against approaches working with abundance values for the search of indicators.

Indicators along the Species Number Gradient

The underlying assumption for the presented approach is that, examined for the occurrence of a particular species, the distribution of species numbers shows an accumulation in a limited range. Especially for indicator species, the distribution should form clusters along the gradient of increasing species number. Biodiversity is a function of environmental variables, and numerous studies have described and used the concentration of the distribution of plants along environmental gradients (e.g., Curtis 1959; Ellenberg et al. 1992; Diekmann 2003). Otherwise, it has to be mentioned that species richness and ecological gradients are not always simply correlated. According to the intermediate disturbance hypothesis (Grime 1973; Connell 1978; Hobbs & Huenneke 1992), species numbers are highest in stands with a moderate disturbance level compared to those with a very low or very high level of disturbance. In interpreting results of our approach, this fact has to be considered.

As the analysis of our data shows (Fig. 3), the concentration of the distribution seems to be approved for the flora of managed forests on acidic sites in central European low mountain ranges, justifying the use of our approach.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Literature Cited

All over the world, there are attempts to convert anthropogenic forests into more natural ones (e.g., Parrotta et al. 1997; Olsthoorn et al. 1999; Zerbe 2002). As Noss (1999) states, biodiversity assessment is also necessary for managed forests especially due to the lack of natural forests in central Europe and the observed occurrence of endangered species. The indicators identified by the present study are a useful device for a rapid inventory. For the assessment of ecosystems as largely implemented in the context of restoration ecology, Landres et al. (1988) recommend to incorporate the use of indicators into a comprehensive risk analysis that focuses on species and key habitats and therefore combines several indicators.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Literature Cited

The study has been financially supported by the German Federal Ministry of Education and Research within the research program BIOTEAM (Fkz. 01 LM 0207). The manuscript was significantly improved by instructive comments from W. Seidling and M. Manthey. We are indebted to P. Herreid for proofreading the manuscript.

Literature Cited

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Literature Cited
  • Andow, D. A. 1991. Vegetational diversity and arthropod population response. Annual Review of Entomology 36: 561568.
  • Barthlott, W., Mutke, J., Braun, G., and G. Kier. 2000. Die ungleiche globale Verteilung pflanzlicher Artenvielfalt—Ursachen und Konsequenzen. Berichte der Reinhold-Tüxen-Gesellschaft 12: 6784.
  • Beierkuhnlein, C. 1998. Biodiversität und Raum. Die Erde 128: 81101.
  • Belbin, L., and C. McDonald. 1993. Comparing three classification strategies for use in ecology. Journal of Vegetation Science 4: 341348.
  • Bemmerlein-Lux, F. A., Fischer, H. S., and R. Lindacher. 1994. Umwandlung von Artmächtigkeitsskalen und Bedeutung skalarer Transformation in der Vegetationskunde. Hoppea 55: 645656.
  • Braun-Blanquet, J. 1964. Pflanzensoziologie. Springer-Verlag, New York.
  • BSBI (Botanical Society of the British Isles). 2004. BSBI database. Botanical Society of the British Isles, Cardiff, United Kingdom (available from assessed August 2, 2004.
  • Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18: 117143.
  • Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199: 13021310.
  • Cook, S. E. K. 1976. Quest for and index of community structure sensitive to water pollution. Environmental Pollution 11: 269288.
  • Curtis, J. T. 1959. The vegetation of Wisconsin. The University of Wisconsin Press, Madison.
  • Deutscher Wetterdienst. 1964. Klimaatlas von Niedersachsen. Deutscher Wetterdienst, Offenbach a.M., Germany.
  • Diekmann, M. 2003. Species indicator values as an important tool in applied plant ecology—a review. Basic and Applied Ecology 4: 493506.
  • Dierßen, K., and K. Kiehl. 2000. Theoretische Grundlagen zur Definition, Messung und Bedeutung von Diversität. Pages 721 in F.Klingenstein and R.Wingender, editors. Erfassung und Schutz der genetischen Vielfalt von Wildpflanzenpopulationen in Deutschland. Bundesamt für Naturschutz, Bonn—Bad Godesberg, Germany.
  • Dirkse, G. M., and F. P. Martakis. 1998. Species density of phanerogams and bryophytes in Dutch forests. Biodiversity and Conservation 7: 147157.
  • Dufrêne, M., and P. Legendre. 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs 67: 345366.
  • Dumortier, M., Butaye, J., Jacquemyn, H., Van Camp, N., Lust, N., and M. Hermy. 2002. Predicting vascular plant species richness of fragmented forests in agricultural landscapes in central Belgium. Forest Ecology and Management 158: 85102.
  • Ellenberg, H. 1996. Vegetation Mitteleuropas mit den Alpen in ökologischer, dynamischer und historischer Sicht. Verlag Eugen Ulmer, Stuttgart, Germany.
  • Ellenberg, H., Mayer, R., and J.Schauermann, editors. 1986. Ökosystemforschung—Ergebnisse des Sollingprojekts. Verlag Eugen Ulmer, Stuttgart, Germany.
  • Ellenberg, H., Weber, H. E., Düll, R., Wirth, V., Werner, W., and D. Paulißen. 1992. Zeigerwerte von Pflanzen in Mitteleuropa. Erich Goltze KG, Göttingen, Germany.
  • Frahm, J.-P., and W. Frey. 2004. Moosflora. Verlag Eugen Ulmer, Stuttgart, Germany.
  • Gaston, K. J. 1992. Regional numbers of insects and plant species. Functional Ecology 6: 243247.
  • Gerlach, A. 1970. Wald- und Forstgesellschaften im Solling. Schriftenreihe für Vegetationskunde 5: 7998.
  • Grime, J. P. 1973. Competitive exclusion in herbaceous vegetation. Nature 242: 344347.
  • Haeupler, H. 2000. Biodiversität in Zeit und Raum—Dynamik oder Konstanz? Berichte der Reinhold-Tüxen-Gesellschaft 12: 113129.
  • Heywood, V. H., R. T. Watson, Baste, I., and K. A. Gardner. 1995. Introduction. Pages 119 in V. H.Heywood, R. T.Watson, and I.Baste, editors. Global biodiversity assessment. Cambridge University Press, Cambridge, United Kingdom.
  • Hill, M. O. 1979. TWINSPAN—a FORTRAN program for arranging multivariate data in ordered two-way table by classification of the individuals and attributes. Section of Ecology and Systematics, Cornell University, Ithaca, New York.
  • Hobbs, R. J., and L. F. Huenneke. 1992. Disturbance, diversity, and invasion—implications for conservations. Conservation Biology 6: 324337.
  • Jackson, D. A. 1997. Compositional data in community ecology: the paradigm or peril of proportions. Ecology 78: 929940.
  • Kratochwil, A. 1999. Biodiversity in ecosystems: some principles. Pages 538 in A.Kratochwil, editor. Biodiversity in ecosystems: principles and case studies of different complexity levels. Kluwer Academic Publishers, Dordrecht, the Netherlands.
  • Landres, P. B., Verner, J., and J. W. Thomas. 1988. Ecological uses of vertebrate indicator species: a critique. Conservation Biology 2: 316328.
  • Levin, S. A. 1997. Biodiversity: interfacing populations and ecosystems. Pages 277288 in T.Abe, S. A.Levin, and M.Higashi, editors. Biodiversity: an ecological perspective. Springer-Verlag, New York.
  • LÖBF (Landesanstalt für Ökologie, Bodenordnung und Forsten Nordrhein-Westfalen). 1999. Rote Liste der gefährdeten Pflanzen und Tiere in Nordrhein-Westfalen. Landesanstalt für Ökologie, Bodenordnung und Forsten Nordrhein-Westphalen, Recklinghausen, Germany.
  • Longton, R. E., and S. W. Greene. 1979. Experimental studies of growth and reproduction in the moss Pleurozium schreberi (Brid.) Mitt. Journal of Bryology 10: 321338.
  • Ludwig, G., and M. Schnittler, editors. 1996. Rote Liste gefährdeter Pflanzen Deutschlands. Bundesamt für Naturschutz, Bonn—Bad Godesberg, Germany.
  • Mayer, P., Abs, C., and A. Fischer. 2002. Biodiversität als Kriterium für Bewertungen im Naturschutz—eine Diskussionsanregung. Natur und Landschaft 77: 461463.
  • McGeoch, M. A., and S. L. Chown. 1998. Scaling up the value of bioindicators. Trends in Ecology and Evolution 13: 4647.
  • Mitchell, P. L., and K. J. Kirby. 1989. Ecological effects of forestry practices in long-established woodland and their implications for nature conservation. Oxford Forestry Institute Occasional Papers 39: 1172.
  • Mooney, H. A., J.Lubchenco, R.Dirzo, and O. E.Sala, editors. 1995. Biodiversity and ecosystem functioning: basic principles. Pages 275325 in V. H. Heywood, R. T. Watson, and I. Baste, editors. Global Biodiversity Assessment. Cambridge University Press, Cambridge, United Kingdom.
  • Munn, R. E. 1988. The design of integrated monitoring systems to provide early indications of environmental/ecological changes. Environmental Monitoring and Assessment 11: 203217.
  • NMELF (Niedersächsisches Ministerium für Ernährung, Landwirtschaft und Forsten). 1996. Waldentwicklung Solling—Fachgutachten, Schriftenreihe Waldentwicklung in Niedersachsen 5: 1150.
  • Noss, R. F. 1999. Assessing and monitoring forest biodiversity: a suggested framework and indicators. Forest Ecology and Management 115: 135146.
  • Økland, R. H. 1995. Population biology of the clonal moss Hylocomium splendens in Norwegian boreal spruce forests. I. Demography. Journal of Ecology 83: 697712.
  • Olsthoorn, A. F. M., Bartelink, H. H., Gardiner, J. J., Pretzsch, H., Hekhuis, H. J., and A. Franc. 1999. Management of mixed-species forest: silviculture and economics. IBN Scientific Contributions 15: 1389.
  • Parrotta, J. A., Knowles, O. H., and J. M. Wunderle. 1997. Development of floristic diversity in 10-year-old restoration forests on a bauxite mined site in Amazonia. Forest Ecology and Management 99: 2142.
  • Pärt, T., and B. Söderström. 1999. Conservation value of semi-natural pastures in Sweden: contrasting botanical and avian measures. Conservation Biology 13: 755765.
  • Passarge, H. 1968. Zur Ansprache des natürlichen Nadelholzanteils. Ein Beitrag zur Frage Waldgesellschaft—Forstgesellschaft. Archiv für Forstwesen 17: 1731.
  • Philippi, G., Quinger, B., and O. Sebald. 1993. Die Farn- und Blütenpflanzen Baden-Württembergs, Bd. 2: Spezieller Teil (Spermatophyta, Unterklasse Dilleniidae) Hypericaceae bis Primulaceae. Verlag Eugen Ulmer, Stuttgart, Germany.
  • Sheehan, P. J. 1984. Effects on community and ecosystem structure and dynamics. Pages 5199 in P. J.Sheehan, D. R.Miller, G. C.Butler, and P.Boudreau, editors. Effects of pollutants at the ecosystem level. John Wiley and Sons, New York.
  • Trautmann, W. 1976. Veränderungen der Gehölzflora und Waldvegetation in jüngerer Zeit. Schriftenreihe für Vegetationskunde 10: 91108.
  • USDA/NRCS (United States Department for Agriculture/Natural Resources Conservation Service). 2004. The PLANTS database, version 3.5. National Plant Data Center, Baton Rouge, Louisiana (available from assessed August 2, 2004.
  • Vane-Wright, R. I. 1996. Identifying priorities for the conservation of biodiversity: systematic biological criteria within a socio-political framework. Pages 309344 in K. J.Gaston, editor. Biodiversity: a biology of numbers and difference. Blackwell, Oxford, United Kingdom.
  • Weckesser, M. 2003. Die Bodenvegetation von Buchen-Fichten-Mischbeständen im Solling—Struktur, Diversität und Stoffhaushalt. Cuvillier Verlag, Göttingen, Germany.
  • Wiegleb, G. 2003. Was sollten wir über Biodiversität wissen? Pages 151178 in J.Weimann, A.Hoffmann, and S.Hoffmann, editors. Messung und ökonomische Bewertung von Biodiversität: Mission impossible? Metropolis-Verlag, Marburg, Germany.
  • Wisskirchen, R., and H. Haeupler. 1998. Standardliste der Farn- und Blütenpflanzen Deutschlands. Verlag Eugen Ulmer, Stuttgart, Germany.
  • Woolson, R. F. 1987. Statistical methods for the analysis of biomedical data. John Wiley and Sons, New York.
  • Zerbe, S. 1992. Fichtenforste als Ersatzgesellschaften von Hainsimsen-Buchenwäldern. Vegetationsveränderungen eine Forstökosystems. Berichte des Forschungszentrums Waldökosysteme Reihe A 100: 1173.
  • Zerbe, S. 1999a. Die Wald- und Forstgesellschaften des Spessarts mit Vorschlägen zu deren zukünftigen Entwicklung. Mitteilungen des Naturwissenschaftlichen Museums der Stadt Aschaffenburg 19: 1354.
  • Zerbe, S. 1999b. Konzeptionelle Überlegungen zur zukünftigen Entwicklung von Nadelholzforsten aus vegetationsökologischer Sicht. Archiv für Naturschutz und Landschaftsforschung 37: 285304.
  • Zerbe, S. 2002. Restoration of natural broad-leaved woodland in Central Europe on sites with coniferous forest plantations. Forest Ecology and Management 167: 2742.